Investigation of pulsed actuators for active flow control using
phase locked stereoscopic Particle Image Velocimetry
by
J. Ortmanns(1) and C. J. Kähler(2)
Institut für Strömungsmechanik, TU Braunschweig
Bienroder Weg 3, 38106 Braunschweig, Germany
(1)
E-Mail: j.ortmanns@tu-braunschweig.de
(2)
E-Mail: c.kaehler@tu-braunschweig.de
ABSTRACT
The dynamic behaviour of two pulsed pneumatic actuators were examined to understand first the effect of the system
parameter (frequency, duty-cycle, waveform, pressure, actuator volume and feed line length) on the output signal, and
second, the interaction of the actuator jet with a turbulent boundary layer. The investigation indicates that the skewed
jet becomes most efficient for boundary layer control when the output-velocity has a span-wise component. In this case
a strong large scale secondary flow is generated, due to the asymmetric injection of the fluid, which transfers high momentum fluid from the outer part of the boundary layer in the near wall region. Blowing with a strong component in
stream-wise direction, on the other hand, seems to be less efficient, relative to the oblique one, because of the small
effect in span-wise direction and the momentum loss of the jet due to turbulent mixing and viscosity. In case of the rectangular actuator it could be demonstrated that the efficiency decreases when the orientation of the actuator becomes
perpendicular to the main flow direction. In this case no asymmetric large scale vortex structure is generated which
transfers high momentum fluid from the outer region of the boundary layer towards the wall. Finally, it could be shown
that the momentum pattern of the dynamic jets becomes very similar to the static ones shortly after opening the valve.
However, during the opening and closing of the valve, a strong span-wise modulation of the base flow can be observed
that stimulates the mixing.
1. INTRODUCTION
The control of boundary layer flow separation by means of synthetic jets is a well established method that has been
frequently applied in many fields of fundamental and applied fluid mechanics. The actuation can be done directly, by
transferring momentum into the desired region by means of tangential blowing (Chang, 1976), or indirectly. In this
case, the blockage of the actuator jet is exploited to enhance the mixing of the base flow, similar to mechanical vortex
generators (Johnston and Nishi 1990). In the past many attempts have been made to replace the stationary blowing actuators in favour of dynamic ones (Gad-el-Hak, 2001, McManus et. al. 1996). The motivation for the use of these actuators is the higher efficiency, due to the reduced mass flux, and the excitation of new turbulent structures which promote the mixing, such as starting vortices for instance. Johari and Rixon (2003) observed that with the dynamic actuation an increase of approximately 30 % of peak vorticity and a penetration of 50 % farther into the boundary layer can
be achieved. However, due to the large amount of control parameter required for the operation of these pulsed actuators, many investigations have been performed to optimize the parameter for different applications. Compton and
Johnston (1992) have varied the skew angle of a single round jet with fixed pitch angle at 45 deg and pointed out that a
skew angle of 90 deg is an optimal configuration to produce a strong stream-wise vortex. McManus et.al. (1996)
showed that the ratio between the jet diameter and the boundary layer thickness should be around four for an efficient
control. Nagib et.al. (2001) concluded from the analysis of several experiments that the reduced frequency must be approximately one. In addition they describe that the most effective location for unsteady forcing is near the point of
separation. The influence of the duty-cycle ∆, defined as the opening time of the valve relative to the time of one period, is another operational parameter discussed in the literature, because it leads to a variation of the durability of the
starting vortex ring. In the present investigation, the dynamic behaviour of different pneumatic actuators where examined systematically to examine the physics of the mixing process and the potential of the actuators for flow control ap-
1
plications in the future. Of particular interest was the response of the actuators as a function of the frequency f, duty
cycle ∆, waveform, pressure pV, actuator volume V and feed line length lS. For this investigation a commercial high
speed magnetic valve of low cost, small size and sufficient robustness was applied because by using this device it becomes possible to vary the frequency and amplitude independently. The response of the pulsed actuator to the input
signal of the frequency generator was investigated with a hot-wire probe and a phased-locked stereoscopic and time
resolved PIV system. After the determination of the optimized operation parameter, a flat plate boundary layer experiment was performed to examine the effect of the actuators on the properties of the flow. Although such investigations
are quite challenging due to the strong velocity variations, the small area of the actuator outlets and the difficulties associated with the injection of the particles inside the dynamic actuator, the dynamical behaviour of the actuators can be
determined as will be seen in the following.
2. PRELIMINARY INVESTIGATION
It is obvious that the flow structure of the jet and its response to an input signal depends on the design of the actuator.
Beside the geometry and the size of the exit area S, the volume of the settling chamber V and the length of the feed line
lS are of primary importance. The volume of the settling chamber for example should be large to avoid inhomogeneities due to the piston, but small to avoid a significant damping or modification of the output velocity signal.
For the investigation a modular construction that supports inlets with different size, geometries and orientation and a
continuous variation of the volume V and tube length lS was designed, see figure 1. In the present investigation a round
jet-inlet with a 1 and 2 mm diameter drill and a rectangular one with a variable aspect ratio in the range b = 5 – 20 mm
(width) and h = 0 – 2 mm (height) was studied. A variation of the pitch α and the skew angle β in the range of α = 30
– 90 deg and β = 0 – 180 deg can be adjusted in addition. This becomes important for the investigation outlined in the
next section. The poppet valve is a commercial MH2 fast-switching valve from FESTO with a maximum switching frequency of 150 Hz. It has an operating pressure range from -0.9 to +8 bar and a response time of 2 ms ± 10 %. The feed
pressure is held constant at pV = 1.5 bar at a line position of lp = 3.5 m in front of the valve. The timing control takes
place with a HAMEG 8131-2 function generator which allows to adjust the frequency f and duty cycle ∆. Because of
the valve design – the switch at a defined threshold voltage value – only rectangular input signals can be processed. For
waveform investigations the FESTO MPYE-5 has to be applied, but this device has a maximum frequency of 100 Hz
and an increasing leakage at high frequencies.
Actuator inlet
V
Retainer
Piston
Fig. 1. Actuator concept with a variable adjustlS
Valve
ment of the volume V, the tube length lS
lp
and a flexible change of the jet exit area by
using inlets.
The dynamic behavior of the pulsed pneumatic actuator was investigated with a constant temperature hot-wire anemometer (Dantec CTA 56C01/CTA Bridge 56C17 system with a single probe 55P12). To preclude a mirror effect of
the probe temperature the distance between the hot-wire probe and the jet-orifice was set to z = 1 mm. The valve driving signal of the frequency generator and the probe signal were sampled using a LeCroy Digital Oscilloscope with a
sample rate of 104 samples/second. A phase locked mean value was calculated based on 100 sweeps.
2
a)
b)
120
120
V = 500 mm3
3
V = 2000 mm 3
V = 10000 mm
100
100
80
80
w [m/s]
80
60
w [m/s]
f = 10 Hz
f = 50 Hz
f = 100 Hz
f = 150 Hz
100
w [m/s]
c)
120
60
60
40
40
40
20
20
20
0
0
0
0.25
0.5
0.75
1
k = f ⋅ T [-]
lS = 150 mm
lS = 400 mm
lS = 1000 mm
0
0
0.25
0.5
0.75
1
0
0.25
k = f ⋅ t [-]
0.5
0.75
1
k = f ⋅ T [-]
Fig. 2. Hot-wire measurements: Dynamic behaviour of a round actuator (d = 2 mm),
pV = 1.5 bar, 1 mm above the exit.
a)
Jet velocity at different actuator frequencies f, V = 2000 mm3, lS = 1000 mm
b)
Jet velocity at different settling chamber volumes V, lS = 1000mm, f = 100 Hz
c)
Jet velocity at different tube length lS, V = 2000 mm3, f = 100 Hz
Figure 2 indicates for a rectangular input signal the effect of the frequency f, the volume V below the actuator orifice
and the length of the tube lS on the velocity signal above the actuator drill. It can be seen that with an increasing frequency the normalized time delay of the output signal increases, according to figure 2a). This effect is reinforced by an
increasing volume of the settling chamber, see figure 2b. In addition it can be seen that an oscillation occurs after closing of the valve. This effect, which has been already observed by McManus et. al. (1995) is associated with the resonance in the tube volume between the valve and the jet orifice. The oscillation frequency fN is given by the tube length
according to the following equation with a being the speed
of sound in air.
4000
3500
a
= f (l S )
lS
3000
This equation is displayed in figure 3 (solid line) and the
dots represent the frequency calculated from the results in
figure 2a). Figure 2c) shows finally, that the maximum value
of the jet velocity decreases with increasing tube length due
to the increasing drag. The results indicate that the rectangular input signal is best reproduced by the actuator when
V = 2000 mm3, lS = 1000 mm and pV = 1.5 bar. Based on
these results it was decided to use these settings for the
following investigations. The frequency was set to f = 100
Hz in order to obtain a rectangular signal without the
oscillation and leackage flow after closing the valve and the
duty-cycle is 50 %. Larger duty-cycles promote the leackage
flow when the actuator is closed and decrease the efficiency
of the actuator.
To validate the properties of the jet flow in air at rest, the
Stereoscopic Particle Image Velocimetry (3C2D-PIV) dis-
3
2500
fN [Hz]
fN ~
2000
1500
1000
500
0
0
200
400
600
lS [mm]
800
1000
1200
Fig. 3. Oscillation frequency as a function of the feed
tube length lS according to the equation. The
dots represent the result from the hot-wire
measurements.
played in figure 4 was installed. The illumination was performed with
a frequency doubled Nd:YAG double-pulse laser system (Quantel
Brilliant) with the following properties: Pulse energy 50 mJ at λ = 532
nm; Beam-pointing stability ± 5 %; Beam-divergence 0.43 mrad; Direction stability < 100 µrad; Pulse-width 4.2 ns. The laser beam was
formed into a light-sheet, approximately 1 mm in thickness, to illuminate the tracer particles which were generated with a 12 hole-nozzle
atomizer. The light scattered by the particles was recorded by means
of two Peltier cooled CCD cameras (FlowMaster 3S) equipped with
TAMRON 180/3.5 Macro lenses.
1
The observation distance was approximately 1030 mm with an opening angle of 86 deg between both cameras and the field of view was
46.1 × 28.0 mm2. The evaluation of the recordings was achieved with
a 2nd order accurate multi-grid method, developed by LaVision, with
32 × 32 pixel interrogation windows and 50 % overlap. For the investigation the excitation frequency was sampled at eight phase angles
(∆ϕ = 45 deg) whereby ϕ = 0 deg corresponds to w = 0.9 wmax in all
cases. To avoid a velocity bias due to loss of seeding when approach2
ing the jet area, a special overpressure seeding atomizer was designed,
that was installed in the feed line of the actuator system. The design is
based on a standard seeding atomizer. Additionally, the pressure from
Fig. 4. Experimental setup Stereoscopic
the supply system is present over the liquid surface by a bye-pass line
Particle Image Velocimetry
and can be adjusted to generate a pressure difference relative over the
seeding nozzle. The pressure drop over a typical four-hole nozzle is
1) Actuator experiment
limited and thereby a seeded air supply for the actuator can be realized. The seeding material for the pressure tube was DEHS with a
2) Overpressure seeding atomizer
nominal particle size of 1 µm. A detraction of the valve operation
3) 3D traverse.
could not be observed. Oil particles generated with a 12-hole nozzle
atomizer were used to seed the wind tunnel flow. The rectangular
master signal is provided from the frequency generator HAMEG 8131-2 with the appropriate actuator frequency f and
the duty-cycle ∆. The signal is connected with the PIV-PC by an external trigger input of the programmable timer unit
(PTU), which is a 16 channel sequencer-card with integrated phase-shifter (Hardsoft PTU 8). This allows an adjustment of a delay in the range of approximate 0 – 40 ms. So according to the results of the hot-wire measurements a
phase shift between the input signal and the actuator response can be set. Unfortunately this is only possible for actuator frequencies above 25 Hz. For frequencies below 25 Hz, the input signal was phase shifted by half a period using an
inverter. One of the main problems with the determination of the three dimensional flow field of the actuator lies in the
dynamic of the out of plane component and the high velocity gradient between the jet and the steady air volume. In
order to capture the illuminated particle structures and to prevent a high level in loss of pairs the time delay between the
illumination has set to ∆t = 2 µs because of the high jet velocity at the exit. This results in a dynamic range of 0.1 < ∆x
< 4 pixel. Based on the assumption that the resolution of the evaluation method is around 0.05 pixel, the ratio between
the largest displacement to the measurement error is about 40. For the experiments described in section 3 this restriction
was not a problem.
3
Figure 5 shows the flow structure of a round (left) and rectangular (right) pulsed jet in a steady environment for
pV = 1.5 bar, V = 2000 mm3, lS = 1000 mm, S = π mm2, ϕ = 0 deg, f = 100 Hz and ∆ = 50 %. Clearly visible are the
a)
b)
25
25
20
20
15
15
Fig. 5. Comparison of two different jet orifices with identical area (S = π mm2) at
ϕ = 0 deg, f = 100 Hz:
z [mm]
z [mm]
10
10
5
5
0
-15
0
-15
a) round, d = 2 mm
b) rect. b = 10 mm,
-10
-5
0
x [mm]
5
10
15
-10
-5
0
x [mm]
4
5
10
15
h = 0.314 mm
symmetric vortex structures, which are sections of the starting vortex, and the entrainment process in the wake of this
vortex ring. A temporal sequence indicate that after closing the valve this flow structure moves away from the wall and
the amplitude decreases gradually. The general flow characteristics show the same behaviour for all frequencies investigated here (f = 10, 50, 100 and 150 Hz). Basically, the difference between the maximum jet velocity (compare figure
5a and b)) is a result of the losses due to the different length of the perimeter between both geometries by a factor of
3.28. The low axial velocity at the wall is a spatial averaging effect, because of the light sheet thickness of about 1 mm
whereas the slot height has a dimension of h = 0.314 mm. An error may occur if the captured particle images represent
a mean value of the fluid motion in the light sheet.
3. TURBULENT BOUNDARY LAYER CONTROL
To examine the effect of the orientation, the geometrical properties and the operation parameter of the actuators on the
flow structure in a turbulent boundary layer, the actuator was installed 300 mm behind the elliptical leading edge of a
flat plate, see figure 6. To generate a two-dimensional flow, the flat plate was mounted horizontally between a pair of
side walls. In a previous experiment the flow quality along the flat plate was examined in detail to validate the boundary conditions (Kähler and Scholz, 2003). However, here the natural flow was disturbed by using a trip wire installed
40 mm behind the leading edge of the plate. The boundary layer thickness was about δ ≈ 14 mm at the jet position at
Turbula-
Flat
Width: 248 mm
Fig. 6. Flat plate experiment: Acryl glass plate with the actuator installed flush with the surface.
14 m/s free stream velocity. The experiment was performed in the open test section of the small continuous atmospheric
wind-tunnel of the Institute of Fluid Mechanics located at the Technical University of Braunschweig. The facility is
designed in a sheet metal construction. The basic dimensions of the tunnel are 4997 mm × 2450 mm, with a nozzle diameter of 505 mm and a 940 mm long test section. The settling chamber is equipped with a flow straightener and a
steel screen. The maximum flow velocity is 65 m/s and the turbulence level at the center of the jet cross-section and
200 mm behind the nozzle exit was determined to 0.85 % at 40 m/s free stream velocity. For the PIV experiments, the
free stream velocity was U = 14 m/s. The orientation of the light-sheet plane was perpendicular to the main flow (yzplane) and the stream-wise location was x = 10, 25, 40 and 60 mm according to figure 7 and 8. The observation distance was approximately 650 mm and the angle between the two cameras in stereoscopic recording configuration was
set to 90 deg. The field of view measures 50.1 × 33.9 mm2 with TAMRON 180/3.5 Macro lenses. The evaluation of the
recordings was achieved with 16 × 16 pixel interrogation windows and 50 % overlap and the resultant spatial resolution
was 0.3 mm. In the following each displayed velocity field represents an average of 100 statistical independent measurements.
a)
b)
Fig. 7. Results in the case of static blowing for comparison with figure 8.
a) Round actuator, d = 1 mm, α = 45 deg, β = 60 deg
b) Rectangular actuator, b = 10 mm, h = 0.314 mm, α = 90 deg, β = 60 deg
5
a)
d)
b)
e)
c)
f)
Fig. 8. Distribution of the mean momentum variation relative to the undisturbed flow (without operation of
the actuator) for the round jet (left) and the rectangular jet (right). The measurements were performed
in four planes (x = 10, 25, 40, 60 mm) behind the actuator at f = 100 Hz, ∆ = 50 %, ϕ = 180 deg.
a) Round actuator d = 1 mm, α = 90 deg, β = 90 deg
b) Round actuator d = 1 mm, α = 45 deg, β = 90 deg
c) Round actuator d = 1 mm, α = 45 deg, β = 60 deg
d) Rectangular actuator, b = 10 mm, h = 0.314 mm, α = 90 deg, β = 90 deg
e) Rectangular actuator, b = 10 mm, h = 0.314 mm, α = 90 deg, β = 60 deg
f) Rectangular actuator, b = 10 mm, h = 0.314 mm, α = 90 deg, β = 30 deg
Figure 7 shows for comparison the static results for the round jet with α = 45 deg and β = 60 deg and the rectangular jet
with β = 60 deg. Blue indicates a loss of momentum relative to the undisturbed flow and red represents an increase of
stream-wise velocity. The spatial dimension of the vortex structure in the steady case is large but farther away from the
wall. In addition it can be seen that in the steady case a strong vertical flow is present in the near wall region but also a
strong upward motion above the vortex structures. This may be a result of a blockage effect induced by the jet.
In the dynamic case, shown in figure 8 for the round (left) and rectangular (right) actuator, it can be seen that the vortex
centres are much closer to each other and the structure is more compact. A comparison of the dynamic excitation indi-
6
cates that in any state of the dynamic process the upwash flow is as massive as in the steady case. The
increasing and decreasing of the dominance of the
induced vortex structures are conducted by a changing blockage effect.
a)
The sequence of results displayed in figure 8 a-c)
reveals for a fixed phase angle of ϕ = 180 deg the
distribution of the mean momentum variation relative
to the natural flow without operation of the actuator.
The upper result was measured behind the actuator
with the round orifice. The orientation of the jet axis
was parallel to the z direction which is normal to the
wall. In case of the centre result the jet axis was in the
yz plane (stream-wise wall-normal) with α = 45 deg
and in the lower result the axis has a steam-wise velocity component against the main flow direction in
addition with α = 45 deg and β = 60 deg. It can be
seen clearly that without a span-wise jet component
no high momentum fluid is transferred in the near
wall region. However, when the jet axis is rotated
around the z axis the flow around the jet becomes
asymmetric and high momentum fluid from the outer
region of the boundary layer is transferred indirectly
towards the wall. This result indicates nicely the sensitivity of the actuator performance on the control
parameter.
b)
Figure 8 d-f) shows the results for the rectangular
actuator, with b = 10 mm and h = 0.314 mm. In the
upper image the long side of the orifice was parallel
to the y axis (skew angle β = 90 deg). The distribution
of the mean momentum variation relative to the natural flow without operation of the actuator indicates
that far away from the wall the stream-wise momentum is enhanced but in the near wall region, right
behind the jet, the flow is delayed. This implies that
this actuator concept is unable to prevent flow separation. However, when β is altered to 60 and 30 deg, as
shown in the other two images of the same figure, an
efficient transfer of stream-wise momentum towards
the wall takes place due to a spiral like secondary
motion of the base flow. As this spiral motion is interacting with the starting vortex structures this interaction reinforces the mixing process. The vector field
shown in figure 9 illustrates the vertical motion in
more detail. It was measured 10 mm behind the actuator at the phase angle ϕ = 180 deg for the skew angle
β = 90, 60 and 30 deg (top to bottom).
c)
Fig. 9. Velocity distribution with the out of plane
component at ϕ = 180 deg. Rect. actuator
b = 10 mm, h = 0,314 mm, f = 100 Hz,
4. SUMMARY
∆ = 50 %.
a)
α = 90 deg, β = 90 deg
The dynamic behaviour of different pneumatic actuators were examined to determine their potential for
flow control applications in the future. Of particular
interest was the response of the actuators as a function of the frequency, duty-cycle, waveform, pressure,
actuator volume and the feed line length, but also
b) α = 90 deg, β = 60 deg
c)
α = 90 deg, β = 30 deg
7
effects of the geometrical properties of the different actuators were examined. However, due to the strong velocity
variations and the small area of the actuator outlets only the far field structure of the jet could be investigated. The investigation indicates that the skewed jet becomes most efficient when the jet axis is not aligned with the main flow direction, because in this case a strong large scale secondary flow is generated that transfers high momentum fluid in the
near wall region. Blowing with a strong tangential component, on the other hand, seems to be less efficient, relative to
the oblique one, because of two reasons. First, no spiral motion of the base flow is induced due to the symmetry. Second, the decrease of the jet velocity is to strong for an efficient momentum transfer in the near wall region. In case of
the rectangular actuator it could be demonstrated that the efficiency decreases when the orientation of the actuator becomes perpendicular to the main flow direction, because in this case no asymmetric large scale vortex structure is generated that transfers high momentum fluid from the outer region of the boundary layer towards the wall. Finally, it
could be shown that the momentum pattern of the dynamic jet becomes very similar to the static one shortly after opening the valve. However, during the opening and closing of the valve, a strong span-wise modulation of the base flow
can be observed that stimulates the mixing.
REFERENCES
Chang P. K. (1976). “Control of Flow Separation.”
Hemisphere Publishing Corporation, McGraw-Hill Book Company.
Compton, D. A. and Johnston, J. P. (1992). “Streamwise Vortex Production by Pitched and Skewed Jets in a Turbulent
Boundary Layer.”
AIAA Journal, Vol. 30, No. 3, pp. 640-647.
Gad-el-Hak, M. (2001). “Flow Control: The Future.”
Journal of Aircraft, Vol. 38, No. 3, pp. 402-418.
Johari, H. and Rixon, G. S. (2003). “Effects of Pulsing on the Vortex Generator Jet.”
AIAA Journal, Vol. 41, No. 12, pp. 2309-2315.
Johari, H., Pacheco-Tougas M. and Hermanson, J. C. (1999). “Penetration and Mixing of Fully Modulated Turbulent
Jets in Crossflow.”
AIAA Journal, Vol. 37, No. 7, pp. 842-850.
Johnston, J. P. and Nishi, M. (1990). “Vortex Generator Jets – Means for Flow Separation Control.”
AIAA Journal, Vol. 26, No. 6, pp. 989-994.
Kähler, C.J. and Scholz, U. (2003). „Investigation of laser-induced flow structures with time-resolved PIV, BOS and IR
technology.”
5th International Symposium on Particle Image Velocimety, Busan, Korea, September 22-24, Paper 3223.
McManus, K., Ducharme, A., Goldey, C. and Magill, J. (1996). “Pulsed Jet Actuators for Suppressing Flow Separation.”
AIAA Paper 96-0442.
McManus, K. R., Joshi, P. B., Legner, H. H. and Davis S. J. (1995). “Active Control of Aerodynamic Stall using
Pulsed Jet Actuators.”
AIAA Paper 95-2187.
Nagib, H., Kiedaisch, J., Greenblatt D., Wygnanski, I. and Hassan, A. (2001). „Effective Flow Control for Rotorcraft
Applications at Flight Mach Number.“
AIAA Paper 2001-2974.
8